Three Expeditions in One Day: The Practice Restarted, and the Interference, Measured
Version 1.2 — October 27, 2026. Version 1.0 was drafted earlier the same day, before the day's audio files had been correlated; 1.1 folded the correlation in. This version adds the day's photographs — their metadata, their timeline, and the light they were taken in — and upgrades one placement on the strength of them.
Today I went out three times. That was not the plan; it is simply what the day became — a short dawn expedition, a short riverside Soundwalk late in the morning, and a walk to the library in the afternoon that doubled as an experiment. Between and after the outings came the other half of the work, the half that happens at the desk: listening, measuring, comparing spectra. Field Art does not just happen in the field. It also happens in the studio and in the labs — the art studio, the sound design and research laboratory. I go out and collect data, and I bring it back to the Historiotheque and turn it into cultural artifacts. Today was the whole circuit in a single day, and this post is its log: three field logs, and then the investigation they fed.
One decision governed all three expeditions. I am restarting my Field Art practice from scratch — deliberately. A few methods of data capture at first, done right, in the right order, and the others added later, one at a time. So today there were no screenshots and no video. Recordings, photographs, a Track, and — new today — the first sensor data. The weather I did not capture in the field at all; it can be recovered afterwards, and some of it I simply record the way it felt on the skin, which is also data, of the phenomenological kind.
Field Log 002 — October 27: The Blue Hour Expedition, and an Experiment at the Picnic Table
Date: Tuesday, October 27, 2026 (all times EDT, GMT-04:00). Out just before 7:00 a.m.; home at exactly 7:45 a.m.
Waypoints for Field Recordings: the park and the Pavillon Isaac-Vandandaigue (W3); the Quai du parc Saint-Henri (W1), where the expedition ended.
Kind of work: Audio field recording and photography, with one Organic Maps Track.
Note on this log: written the same day, from my reports; the audio files were uploaded and correlated the same evening, and the recordings are tabled at the end of this log. As established for Field Log 001, a file's timestamp is the moment the recording stopped; the start is the timestamp minus the duration.
The tone of the day
I opened the day at 6:00 a.m. with the recording I have come to think of as the tone of the day: about two minutes of Atmospheric Tone, the ambient sound outside, taken before anything else. (The file, 261027_0249.wav, runs 06:00:03–06:02:08. My own log of the morning said 6:04 — the recorder's clock is the one that counts, and the difference is recorded here rather than smoothed over.) A recording studio takes room tone to know the room it is working in. This is the same gesture, made at the scale of a day — except that the recording also carries, or will come to carry, the day's other data with it. It is a super important practice, and it now opens every expedition day.
The Conference
The moment I stepped outside there was a massive murder of crows, seen and heard, already moving off. I recorded immediately, but by the time I pressed record they were much further away — the recording, 261027_0250.wav, runs 06:54:40–06:55:28, and what it holds is a Conference already in departure. The same thing happened yesterday morning, and I now know what I was looking at, both times. In late autumn, American Crows abandon their small breeding territories and sleep together in communal roosts — sometimes thousands of birds, often near water and mature trees, which puts the Old Village, on the Richelieu's bank, in exactly the right geography. What fills the sky at this hour is the roost breaking up: the dawn dispersal of a community that has slept together for safety and warmth, scattering into the day.
Before the flock disperses into its foraging groups, the birds hold a staging at high points near the roost's edge — rooftops, treetops, wires — loud with caws. It is a roll call and a coordination: an exchange of information about where food was found the day before, the spring's young birds learning the flock's dynamics from the inside, and, not least, a safety audit. Great Horned Owls hunt roosting crows in the pre-dawn hours, when crows, nearly blind at night, cannot defend themselves; the dawn uproar is partly the flock establishing that the coast is clear before anyone drops to the vulnerable ground. There is an acoustic logic to the hour, too: at first light the air is still, its turbulence low, and sound carries — the calls reach the whole congregation at once, through the same quiet medium my tone of the day had sampled an hour earlier. So my word was better than I knew: corvid researchers describe these gatherings in terms not far from a Conference. The crows dispersed into the pink light I had come out for, and I went on up the hill.
The light
I timed the outing for the light, and the visual data matters as much as the audio — I want that on the record plainly. Sunrise today was at 7:25 a.m.; the Blue Hour, I had been told, begins some twenty-five to forty minutes before sunrise, and the expedition was built to stand inside it. Towards the west — the direction of the church of Saint-Mathieu-de-Beloeil, across the river — the low sky held a rich pinkness, and with it bright, light blues: ethereal and magnificent, the church set against it, and a moon above the church. Against that luminous west, the Old Village stood in contrast: the east-facing surfaces of its architecture in deep, dark shadow hues, and the autumn foliage everywhere completing the chord. Pink, shadow, foliage: that is the morning's visual field, and the photographs carry it. By the time I came down to the dock at the end, the sun had been up twenty minutes and the pink was already past its peak — a few minutes of colour is the whole event, and you have to be out in it, on purpose, at exactly the right time. I was.
The subtleties deserve their own sentences, because they are the kind that evaporate. The morning's colour is faded, as though in a dreamy haze — nothing at full strength, everything veiled and tender. The old building at the corner of St-Anne and St-Hippolyte stands in dark shadow, facing westward into the last of the pink. And the sight facing west, walking back down St-Anne street later in the descent, is the counterpart view: the street falling away toward the river light. In the Daylight Series I keep returning to the effects of light in terms of colour temperature, and this morning is a study in the cool end of it — a low, blue-pink temperature over everything, against which the few warm notes (a lit window, the foliage) read almost hot. There is a sublime poetry to it that the data does not exhaust, and the log should say so without embarrassment: the poetry is part of what was there. Later in the day, as Log 003 records, the same village under a high sun is another world entirely — blues so saturated you wonder if they are even real. The morning is the subtler instrument. I would not trade it.
The experiment at the picnic table
The rest of the expedition was given to the interference — the distortion found in Sunday's park recordings, which I have named The Interference in The Park Event. Today's test was a walk-away. At the picnic table in the park, the exact spot and direction of Sunday afternoon's recordings, I recorded a baseline. Then I walked away from the table on both axes, in both directions, counting my steps aloud at intervals of five, to about twenty-five steps — at my stride, at least twenty-five yards — making still recordings at the marks, and counting my way back to the table each time. Finally I walked, recording, toward the cell tower, which is also the direction of the town hall garage.
Listening afterwards on the recorder's own speakers — I could not make out fine detail, but this much was unmistakable: there was tons of distortion in the recordings, the same exact interference pattern, at the table, at twenty-five yards from it in every direction, and on the walk toward the tower. No edge, no fading-out, within the radius I tested. The zone is bigger than the bench's immediate vicinity, and after today it has a measured size of at least that much.
The close
I came back down the hill photographing throughout — dozens of photographs in all, from the Blue Hour until I got home — and finished on the dock, W1, with final photographs of the church of Saint-Mathieu-de-Beloeil across the river. The sky's pink hue, like Sunday morning, was a few minutes past its peak. Home at 7:45. It was a great field expedition, and I felt really great.
The recordings, correlated
Eight files, 13 minutes 10 seconds. The exact assignment of the middle files to the legs of the walk-away experiment is deliberately not guessed here: the counting is audible in the recordings, and the files will be identified by listening, not by arithmetic alone.
The photographs, correlated
The morning's expedition produced dozens of photographs; what follows is a sample of them — the frames I have pulled out and examined so far — in the order the expedition lived them. All photographs in the sample: Samsung Galaxy A16 5G, f/1.8.
The camera's own numbers, in this sample, trace the dawn as faithfully as any sentence here: ISO 400 at 7:13, when the light was barely workable; ISO 25 by 7:40, the sensor drinking full daylight. Between those two readings, the Blue Hour opened and closed. The metadata is a light log, kept by the instrument, whether or not I remembered to keep one myself — and the rest of the morning's frames, still unexamined, are in the archive for the image work to come.
Field Log 003 — October 27: A Short Soundwalk at the Riverside
Date: Tuesday, October 27, 2026, late morning; back by about 11:30 a.m.
Waypoints for Field Recordings: the dock at the bottom of St-Anne street, in the Old Village — a waypoint I created today in Organic Maps. Whether it coincides with W1, the Quai du parc Saint-Henri, will be settled from its coordinates, not assumed.
Kind of work: A Soundwalk — photography, audio and sensor data logging. A limited run, deliberately.
Cold water, warm street
It was very humid and cold by the river, and sunny. Down at the dock it was genuinely cold; up on St-Anne street, in the sun, it felt warm enough. That contrast — the same morning, a few hundred metres apart, two different climates on the skin — is exactly the kind of scalar fact these logs exist to keep. I photographed the church of Saint-Mathieu-de-Beloeil across the Richelieu from the dock, and, coming up, the scene up St-Anne street with the mountain in the background. It was lovely.
At the dock I also ran the first sensor logging of the revamped practice: with Sensor Logger, a recording of the barometric pressure, and of the light, with a light-meter type reading; and a capture of the noise with NoiseCapture. First uses, all of them — the point today was not mastery but correct beginnings.
And the light had changed its nature entirely. At 10:32 I photographed the stone church here in the village (20261027_103214.jpg, 1/476, ISO 25): its east-facing stonework blazing in full sun out of the east, the sky a blue of an entirely different order from the morning's. At 10:40, from the dock, the church across the river again (20261027_104014.jpg, 1/658, ISO 25) — the same spire I had photographed at 7:41 in pink and shadow, now standing in deep, saturated blues of sky and water so intense you wonder if they are even real. Everything after noon is like that today: bright, clear, distinguished, each colour at full strength. Set the two church photographs side by side — 07:41 and 10:40 — and the whole argument of the day is in them: one subject, two colour temperatures, a dream and a declaration. The morning is subtler. The midday is surer. The practice needs both, and today it happened to collect them three hours apart.
The human notes
A toasted bagel with cream cheese at the local shop. Then a bottle of spring water at the corner store, and a few minutes' talk with the owner about the morning expedition, and about the traffic on the river road — which was heavy enough that I could not safely cross it without unnecessarily putting my life in danger. Fieldwork in a village includes the village. It always has.
I will be doing short Soundwalks like this one more and more, alongside the longer expeditions I have always done, weather permitting. The short run is not a lesser form. It is the form that can happen almost any day — and a practice that can happen almost any day is the one that accumulates.
One recording falls in this outing's window: 261027_0257.wav, 10:42:55–10:43:44, forty-nine seconds. Version 1.1 of this post left its place unclaimed. The photographs now place it, by correlation and with high confidence: it begins two minutes after the 10:40 photograph of the church, taken from the dock, in the middle of the dock session where the Sensor Logger and NoiseCapture work was also done. The recording is the dock session's audio. It is stated here as a correlation, and will be confirmed by listening — but the circumstantial case is complete: right place, right minutes, no other candidate. The soundscape record of the Soundwalk proper is this one short file, which suits a limited run.
Field Log 004 — October 27: The Walk to the Library — An Onset, Located
Date: Tuesday, October 27, 2026, afternoon. The recordings run from 12:09 to 12:37 p.m.; the library visit itself — reading, and organizing the physical folders — continued into the afternoon. The afternoon was much warmer than the morning, though I still wore a scarf and a tuque, and my hands were slightly cold.
Waypoints for Field Recordings: from the corner of St-Hippolyte and St-Anne, in the Old Village, up the hill to the park — the picnic table, the Pavillon (W3) — and the Bibliothèque Armand-Cardinal (W4).
Kind of work: A collation experiment: an Organic Maps Track and an audio recording run simultaneously, to be laid over one another afterwards — plus an errand. I had printed two documents in the morning, and I took papers to the library to read and to organize file folders in the physical realm. I never bring my laptop to the library. The desk work waits for home; the library is for paper.
The design
This walk was built to answer one question: travelling toward the park from outside the affected zone, where does the interference start? So I started recording at the corner of St-Hippolyte and St-Anne — audio and Track together — and walked up the hill directly to my picnic table, naming waypoints aloud as I passed them: the elementary school on St-Anne; the corner of Provencher and St-Anne; the train tracks on St-Anne; the entrance to the park; a point halfway from the entrance to the Pavillon Isaac-Vandandaigue; the Pavillon itself; the picnic table. Because the waypoints are spoken inside the recording, whatever the recording contains can be placed. That is the collation: voice, audio and Track, three clocks telling one time.
The distance, corrected twice and then measured
On the way, the investigation corrected itself. I had estimated the cell tower at 300–400 metres from the picnic table. I then estimated it again, by eye, at about as far as I could throw a baseball, maybe a little further. Both were wrong. Measured from coordinates taken in Google Maps — the picnic table at 45.56589824, −73.19072399; the tower, by the bike path, at 45.56486921, −73.19256398 — the distance is about 180 metres, the tower southwest of the table, with maybe 10–20 metres of uncertainty in the tower's pin. An independent recomputation agrees. Estimates are what measurements replace; that is the whole method in miniature.
What the walk carried
A second recording ran from the picnic table toward the parking lot and the library entrance: the same distinct, piercing frequency and harmonics, the same signature. Inside the library, a recording was clean. At the library entrance, distorted. The zone has an inside and an outside, and the library stands on its edge.
The afternoon's recordings, as the properties give them: six files between 12:09 and 12:37 p.m. — 261027_0258.wav (12:09:07–12:14:50, 5:43), the long file from which the measured sample below was drawn; then 261027_0259.wav (12:22:02–12:22:22), 261027_0260.wav (12:22:42–12:23:04), 261027_0261.wav (12:23:37–12:24:44), 261027_0262.wav (12:24:53–12:25:08), and 261027_0263.wav (12:36:17–12:37:22). Among the short files are the stage, the library entrance and the library interior of the account above; which file is which will be settled by listening, and stated when it is settled. One more file stands apart: 261027_0264.wav, twelve seconds at 14:33:56–14:34:08, later in the afternoon — unassigned for now, and recorded here so that it is not lost.
The Interference, Measured
The rest of the day was the laboratory half: listening, and learning the instruments. I am slowly building the analysis stack — Friture on the Windows laptop, ocenaudio on the Ubuntu machine, Audacity, and, as of today, Raven Lite, which I used years ago in its first version and have now registered again in its second, free of charge. I am also learning to see. My first frequency analysis, in Audacity, was run at an FFT size of 1024 — a bin width of about 47 Hz — and the peak frequencies I extracted by hand scattered accordingly: 1010, 2073, 3004, 3950, 4980… Measured at full resolution, every one of those lines sits within a tenth of a hertz of an exact multiple of 1000. The lesson is now part of the method: the resolution of the instrument is part of the result, and you state it, or you fool yourself with fenceposts.
In Raven Lite, the right track of today's sample shows the thing whole: the waveform swelling and receding in slow waves, and beneath it the spectrogram — horizontal striations, the harmonic comb, crossed by vertical banding, the pulsing structure of the signal. I will make music and soundscapes with these sounds eventually. First, the accounting.
The fingerprint
The same folder of properties held a bonus: Monday's bench tests now have file identities too. The two sessions of Post 002 — phone on airplane mode, phone with data on — are 261026_0245.wav and 261026_0246.wav (15:34:04–15:34:34 and 15:35:14–15:35:54, a consecutive pair) and 261026_0247.wav and 261026_0248.wav (16:01:29–16:02:34 and 16:02:51–16:04:44). Which pair was flown in which mode is not recorded, and is not guessed; the sessions themselves are now files, which is what matters for any future measurement.
Two samples, two days apart — Sunday's from 261025_0242.wav, today's from 261027_0258.wav — measured the same way:
- Fundamental: 1000.01 to 1000.03 Hz in both files, both channels. Identical, across days, to the precision of the measurement.
- The comb: lines at every multiple of 1000 Hz, from 1000 to at least 23000 Hz, their amplitudes decaying in the same profile both days — 2000 Hz a few decibels below the fundamental, 3000 a few more, and so on up the band. What I first reported as "harmonics to 5000 Hz" was only the part of the comb that stood above a first look.
- The notches: the lines at 7000, 14000 and 21000 Hz are essentially absent — 15 to 20 decibels below their neighbours. I had caught this without knowing it: in my own hand-extracted list, the 7000 line was flagged "slightly lower amplitude," and 14000 was missing entirely. Nulls at multiples of 7 kHz are what a pulse train leaves when its pulses are about 1/7000 of a second wide — roughly 0.14 milliseconds per pulse.
- The gating: beneath the main lines, both files carry a fainter comb at every 50 Hz — the signature of a signal repeating every 20 milliseconds. The 1 kHz is not a continuous tone. It is pulsed: pulses of about 0.14 milliseconds, one every millisecond, gated on a 20-millisecond cycle.
- Two doors closed: there are no peaks at 217 Hz or its multiples — the classic GSM cell-phone buzz, which a consulted analysis proposed, is simply not in these recordings. And there is no 60 Hz structure — in a 60 Hz country, ordinary mains hum would grid at 60. A 20-millisecond cycle points instead to electronics that keep digital time, or were built around 50 Hz.
- The envelope: the swelling is real — 13 to 20 decibels of slow rise and fall — and irregular. No strict period. It breathes; it does not tick. The ticking is in the pulses; the breathing is in their strength.
Same fundamental, same profile, same grid, same notches, same right-channel bias, Sunday and Tuesday. One source, fixed, keeping very precise time. A strong regularity. It runs like clockwork — because something in it, literally, is a clock.
The onset, and the shape of the zone
Listening to the library-walk recording, the spoken waypoints did their work. The interference begins at 45.56717700, −73.19153670 — about thirty metres past the train tracks, 269 metres from the tower, 156 metres from the picnic table. From there it amplifies progressively for a short stretch; then it goes intermittent — on and off, fading in and out with gradual decays, never switching like a step function. It comes on strong at the park entrance (45.56645970, −73.19058720; 234 metres from the tower). It is very loud at the halfway point (45.56651978, −73.19094661; 223 metres) — a clearing between coniferous trees that otherwise stand between the point and the tower. From the Pavillon (207 metres) the amplitude skyrockets, up to the picnic table (183 metres), the peak of the zone; and it stays loud through the parking lot to the library entrance, though below that peak.
One anomaly inside the anomaly: a recording made at the stage (45.56620808, −73.19078032; 204 metres from the tower) carries the interference less strongly than the Pavillon — from a point 3.4 metres from the Pavillon point. No source hundreds of metres away does that. Either something very local is radiating within metres of the Pavillon — near fields collapse within metres; that is their signature — or that recording is not comparable to the others, which a back-to-back test will settle.
So the amplitude does not track the tower's distance: strong at 234 metres, weak at 204, maximum at 207. The tower remains a live candidate for the mechanism — the demodulation of a strong radio signal inside an unshielded recorder is real, and 180 metres is close enough to matter. But the map now points into the park itself. The next experiments follow from the map: a shield test at the table — one recorder in a closed steel tin beside one recording in the open — to decide whether the signal arrives as an electromagnetic field or through the microphones; a phone cross-check, with the Physics Toolbox spectrum analyzer and a Sensor Logger session running beside the recorders; an orientation test and a gain test at the same table; magnetometer transects, to interrogate the electrical plant of the park; and a fine amplitude grid around the Pavillon, at metre scale, where the gradient is steepest. The battery is designed. What remains is to run it.
What this day clarified
I have carried the concept of Field Art, and its clusters of concepts, for twenty years, maybe longer — able to talk about it, not always with precision or depth. What clarified it today was not a definition. It was the practice, done analytically, end to end: out into the field with a question, back to the laboratory with data, the data measured, the measurements written. An anomalous acoustic phenomenon appeared in the middle of it — a distortion with an exact signature, strong regularities, a source still unknown — and instead of ruining the recordings, it became the most instructive material I have collected in years. The distortion is going to pay off. It already has: it taught the method faster than a clean day would have.
Part of what I am doing, I now see plainly, is profiling a role: the audio forensics expert, or someone of the sort. I am not claiming that expertise; I am studying what such a person does — how this kind of data is gathered, logged, and worked — so that those data sources and data logs can join the rest of my practice: the atmospheric and phenomenological work, the painting, the research, and in time the music and soundscapes I will build from these very sounds. It sits naturally beside the Deep-Archives Project, which keeps meeting archival failures: here is the opposite event, an artifact that documents itself, relentlessly, at a thousand cycles a second. And it sits inside a longer formation: I am a noise artist at heart. I have studied noise for thirty years and synthesized it, visually and acoustically; the Soundwalks, the acoustic ecology, the dérive — they are all, underneath, the same attention, turned toward the world and listening to what it emits.
All of my projects are coming together in a crescendo, and a dramatic moment. Today, for the first time, I can say I know exactly what Field Art is — because today I did all of it: the field, the laboratory, the log.
Next steps
- Finish the identifications by listening: confirm
261027_0257.wav at the dock; which of the afternoon's short files are the stage, the entrance and the interior; which legs of the walk-away 261027_0254.wav holds. The recordings register and the UCS names for the day (sources FL002, FL003, FL004) follow from that, and the Organic Maps Tracks can then be laid against the recordings they belong to. - Run the experiment battery at the park: the tin test first, then the metre-scale grid around the Pavillon, the magnetometer transects, and the stage/Pavillon back-to-back pair.
- Continue the desk analysis: the amplitude of the tone against the counted steps in the morning's recordings is still to be measured, and the images from the day — the Blue Hour set above all — are next: the image analysis and production phase begins from them.
- Beyond the log itself, the same disciplines are converging in the software work — secondary here, but on the same trajectory: Python, already the instrument of today's analysis, is also what I will use to generate text by randomness and other functions for the Experimental Novel, the novel-as-a-system, with its own noise, fragmentation and contradiction; and the Refcards-System continues as the operation's software strand. One cursus, several instruments. I have a name for the period this belongs to: The Age of Discipline.
— A.G.
